A classification tool to foster self-regulated learning with generative artificial intelligence by applying self-determination theory: a case of ChatGPT

Abstract

Generative AI such as ChatGPT provides an instant and individualized learning environment, and may have the potential to motivate student self-regulated learning (SRL), more effectively than other non-AI technologies. However, the impact of ChatGPT on student motivation, SRL, and needs satisfaction is unclear. Motivation and the SRL process can be explained using self-determination theory (SDT) and the three phases of forethought, performance, and self-reflection, respectively. Accordingly, a Delphi design was employed in this study to determine how ChatGPT-based learning activities satisfy students’ each SDT need, and foster each SRL phase from a teacher perspective. We involved 36 SDT school teachers with extensive expertise in technology enhanced learning to develop a classification tool for learning activities that affect student needs satisfaction and SRL phases using ChatGPT. We collaborated with the teachers in three rounds to investigate and identify the activities, and we revised labels, descriptions, and explanations. The major finding is that a classification tool for 20 learning activities using ChatGPT was developed. The tool suggests how ChatGPT better satisfy SDT-based needs, and fosters the three SRL phrases. This classification tool can assist researchers in replicating, implementing, and integrating successful ChatGPT in education research and development projects. The tool can inspire teachers to modify the activities using generative AI for their own teaching, and inform policymakers on how to develop guidelines for AI in education.